Collections > Electronic Theses and Dissertations > Can a Major National Food Retailer Improve the Nutritional Profile of US Packaged Food Purchases and Reduce Diet-Related Disparities?
pdf

Since 2011, three of the largest national food retailers in the US have implemented healthier food initiatives (HFIs). Given that US children and adults consume the majority of daily energy from food retailers, and that fewer retailers account for an increasingly large share of food purchases, these HFIs have great potential to improve nutrition. These HFIs may especially be important for low-income and race/ethnic minority households, who often face the highest barriers to a healthful diet. However, no independent work has evaluated whether an HFI at a major national food retailer actually improves the nutrient profile of food purchases, whether any changes in nutrient profile are simply the result of attracting more health-conscious customers (i.e. selectivity), or how the selection of a pre-pledge baseline period can affect results. The overarching objective of this research is to develop a methodology to evaluate the impact of the food retailer and food-retailer based HFIs on the nutrient profile of US packaged food purchases (PFPs). Using a dataset of household PFPs, we employ Walmart as a case study to first examine the impact of a single food retailer as a food source in the US and examine who is most likely to shop there. Secondly, using fixed effects models and inverse probability weighting to account for selectivity, we evaluate whether the nutrient profile of Walmart PFPs improved over time, and whether these changes were greater among households eligible for the Supplemental Nutrition Assistance program (SNAP) and race-ethnic minorities. We also examine whether observed shifts were driven by shifts in consumer purchasing (i.e. buying more or less of certain food groups), or driven by product reformulation or introduction of lower-calorie options (i.e. shifts in nutrient profiles within food groups). To test whether improvements in the nutrient profile of Walmart's PFP were attributable to the HFI, we employed counterfactual simulations to compare the observed nutrient profile of post-HFI purchases to the expected trajectory of nutrient profiles based on pre-HFI trends. Finally, we tested whether results varied depending on whether Walmart's stated HFI initiation date (2011) or a data-driven HFI initiation date (2007) was used. We found that not only do the majority of US households buy PFPs at Walmart, the proportion of PFPs purchased from Walmart doubled from 2000 to 2012. Low-income non-Hispanic White and Hispanic households, but not low-income non-Hispanic Black households, were more likely to buy a larger proportion of PFPs at Walmart. From 2000 to 2012, Walmart PFPs showed substantial declines in energy density, sugar density, and sodium density, and these declines were much larger than declines observed in other chain retailers across the same time period. While SNAP-eligible and higher income households showed similar declines over time, race/ethnic disparities in the nutrient profile of Walmart PFPs decreased over time. Hispanics and non-Hispanic Others showing the largest declines in energy density and sodium density. Overall shifts were driven in part by changes in food purchasing, including declines in "less healthy" food groups, like grain-based desserts, candy, and snack foods, as well as by changes in product formulation, as demonstrated by major declines in nutrient densities within food groups. However, we also found that the majority of these shifts occurred prior to 2007, and leveled off in later years. When Walmart's stated HFI initiation date (2011) was used as the cut-point for evaluation, we found that post-HFI shifts in nutrient profile were similar to what would have been expected based on pre-HFI trends. However, when a 2007 HFI initiation date was used, we found that changes in nutrient profile and purchasing patterns were less than expected based on pre-HFI trends. These results indicate that major national food retailers can act as key agents to improve nutrition and potentially reduce diet-related disparities. However, improvements in the nutrient profile of Walmart PFPs appear to be a reflection of secular trends and not necessarily attributable to its HFI. Continued evaluation of HFIs at Walmart and other national retailers is needed in order to understand how retailer-driven changes in food purchasing translate into changes in food intake and impacts on health outcomes, especially for vulnerable sub-populations.